Care and Feeding of the OBC/dc-dc

[JoeS]

Our OBC/dc-dc is proving to be the i-MiEV's Achilles heel.

Not wishing to contaminate kiev's excellent thread detailing the troubleshooting and repair of the On Board Charger (OBC) which also contains the pack-to-12v dc-dc converter, I thought I would start this thread specifically to pursue the minimization of damage to that module.

There are a number of concepts floating around that are perhaps worthy of discussion -

1. 12v battery
2. Charging using L1 (120vac) instead of L2 (240vac)
3. Reducing charging current (by those of us with an adjustable EVSE)
4. Not cutting off input power to the EVSE and instead have the car send a shutoff signal to the EVSE
5. External thermal management; specifically, external cooling fans

I thought we'd start off by having some discussions about the merits of each concept and hopefully be able to zero-in on the specific technical aspect of why it's significant (or not).

Any other concepts to explore that we can add to this list?

 

[PV1]

I'll start with my thoughts on the 12 volt battery.

My white i-MiEV, built February, 2012, recently had the 12 volt battery fail with a shorted cell. Symptoms were very low voltage after storing the car for a couple weeks, and attempting to charge the 12 volt battery showed an error on my battery charger. The 12 volt battery would go to 14.5 volts while charging, though drawing low amps. When the charger stopped to check the battery, resting voltage went down to 11.2 and would not go any higher. Charger threw an error for shorted cell. I replaced the battery before attempting to start the car and had no error lights come on. A 10 mm socket and wrench is the only tool necessary to remove and replace the battery, although a terminal cleaner would help make a clean connection.

This battery was the original 12 volt from the factory and was 7 years old. During my ownership, it has gone dead on several occasions, which I've been able to narrow down to the car drawing excessive current while parked. With some investigative work, I've been able to reproduce the issue consistently. What happens is. I always lock my cars when I park them, so if I need to retrieve an item, I need to unlock the car. Process is:

1. Unlock car with key fob.
2. Open door and retrieve item, not touching any controls or lights in the car.
3. Close door.
4. Re-lock car with key fob.

I've noticed that within 3 days, the 12 volt battery will be discharged, often to the point where the power locks can't unlock the door. This has usually happened over a weekend if I do the unlock/lock procedure above on a Friday night. However, I've been able to leave the car for up to 2 weeks at a time by changing one thing - starting the car. In this case, the process is:

1. Unlock car with key fob.
2. Open door and start car while retrieving item.
3. Leave car run for a few seconds, usually 30 seconds.
4. Turn off car.
5. Close door.
6. Re-lock car with key fob.

Voila. The car can sit for a week and not have the 12 volt battery go dead. So, something wakes up and expects the car to go READY or charge when the alarm is disarmed, but when one of those events doesn't happen the component doesn't return to low power and ends up discharging the 12 volt battery. As a note, this drain does not occur if you have been driving the car and simply park it afterwards. This only happens with a disarm/arm cycle without starting the car in between.

As far as the OBC is concerned, my recommendation would be that we all change out our 12 volt batteries. Since most of the cars on this forum are 2012 models, those batteries are now 7 years old. I'll be replacing the battery in my silver i-MiEV soon. Besides potentially avoiding the charger issue, fresh 12 volt batteries will increase the reliability and reduce weird issues with our cars, which EVs seem to be particularly prone to issues stemming from a weak 12 volt battery. The battery is group size 151R.

 

[PV1]

Concerning items 2, 3, and 4, I normally charge my car at 12 amps on 120 volts (level 1) or 13 amps on 240 volts (level 2). These are the defaults on my EVSEUpgrade cord.

In the summer, I only use level 2 unless I am charging where only level 1 is available. This reduces the active time for the high voltage system and keeps the battery cooler. In the winter, I use level 1 at home unless I need to charge faster in order to keep the battery warmer. The car is parked outside, so everything is usually cold when charging starts.

If I stop charging the car manually at home, it's done through a switched outlet feeding the EVSE and not by pressing the J1772 release button or using the remote. My normal routine is let the car charge fully on its own, and then if I use pre-heating, I stop it with the switched outlet, and then unplug the J1772 connector (switch and EVSE are inside the garage, car is parked outside). I leave my EVSE turned off when not in use. I have been doing this since I got the car in 2013.

I'm posting this information here as a comparison to the lives of the OBCs that have failed. My hope is that this weekend, I can remove the covers of my chargers and check their condition. Both of my cars still charge, but they are parked until we can figure this out.

 

[JoeS]

PV1, thanks for that and for bringing attention to yet another i-MiEV quirk. Never would have suspected that not starting the car after opening the door would leave a higher quiescent drain on the 12v battery.

Continuing with 12v, I'm now seriously toying with installing a "12v" Headway LiFePO4 battery pack, motivated by the significantly-higher quiescent voltage of this pack compared to the OEM lead-acid battery. With the car turned off, this "12v" battery would sit at around 13.3vdc compared with our lead-acid battery which is 12.65vdc at best (room ambient). There's a photo of such a setup in the i-MiEV towards the end of this post (but unfortunately LOWRACER has not been heard from in years):
http://myimiev.com/forum/viewtopic.php?f=9&t=694&p=4035

I recently picked up a cheap LiFePO4 balancing and protection circuit. So far, my testing has revealed that when a cell's upper voltage limit (3.65vdc) is exceeded the voltage will go up a bit more but then a slow sawtooth can be seen. I'll be continuing testing as I want to crank up the overall input voltage above 14.4vdc to see what happens...

I'm not sure that I want to connect this module in series such that both the overvoltage and undervoltage protection is enabled which presumably opens up the connection to the battery. The car would undoubtedly not like this.

Will post whatever I end up doing with 12v.

Regarding externally turning off the input to the EVSE, anecdotally we have a fair number of examples of this being done for years with no issues (knock on wood); for example, I've been controlling the inputs to my two i-MiEV EVSEs using mechanical timers, either when using L1 or L2. For example, http://myimiev.com/forum/viewtopic.php?f=7&t=129&p=2910

Regarding 240vac(L2) vs. 120vac(L1), I've regressed and now primarily use L1. Those failures inside the OBC are sure unnerving... Perhaps someone knowledgeable with dc-dc circuit design can comment on the various component stresses subjected to the different voltage levels.

 

[pbui19]

I too very much appreciate this thread. Some thoughts:

We had noticed early on that the 12v aux batt gets drained if we lock the i using the fob button; so I've gotten into the habit of only to lock with the button on the door. The other source of drainage is the ODLink brick for Canion, so I only plug it in when I am using Canion.

I've also thought about replacing the 12v with LiFe. Given that LiFe tops out at 3.65x4=14.6 v, and the dc/dc tops at 14.2v; I am thinking that it wouldn't need a bms if I first top balance the 4 LiFe cells in parallel a few days before hooking them in series. The problem of course is that the LiFe will live near its top soc all the time.

I have a 2x4 block to prop up the hatch when charging L2 for more than 30 minutes these days.

Is the OBC problem limited to earlier production model ? probably not, as my OBC can get pretty toasty

 

[Don]

After storing our original car for several months, I replaced the 12 volt with an AGM Miata battery and it has worked fine. Need to do that with the second car real soon

Since the charger's failures all seem to be heat related and since it gets hotter here in the summer than most places, I've given up L2 charging completely - It's now used only for our Volt. Neither of my cars have ever been recharged at max wattage, as I've always had my Open EVSE set to 12 amps at 240 volts. Since we began reading about so many failures, I use the OEM EVSE for charging whenever it's hot out - 120 Volts at just 8 amps doesn't heat up the charger that much. The rest of the time I use the EVSE Upgrade on L1 at 12 amps. The way we use the cars, even charging at only 960 watts is no problem at all, so why push things?

Don

 

[PV1]

I think Lithium Titanate cells, wired 6 cells in series, are better suited for aux. battery duty than LiFePO4. Voltage range is 9-16.8 volts, so no worries of the cells sitting at full charge all the time. This is the same chemistry used in the SCiB drive batteries for JDM i-MiEVs and features increased cycle life and higher charge rates.

https://lithium-titanate-battery.com/lto-battery-40120-10000mah-2-4v/

Anyway, regarding my discovery with battery drain, driving the car then parking and locking doesn't cause it with mine, nor does unlocking it and opening/closing a door and leaving it unlocked. It's just when the car is unlocked and re-locked without starting.

 

[JoeS]

This post is about turning OFF power to the car when charging.

There are four ways that the car turns off charging -

1. When it is fully charged, the OBC sends a shutoff signal to the EVSE which stops the power flow

2. When timed using the Remote, presumably the Remote receiver sends a signal to the OBC which then sends the shutoff signal to the EVSE

3. Depressing the pushbutton on the J1772 connector sends the shutoff signal to the EVSE

4. Input power is removed from the EVSE, either by the use of a timer or simply pulling out the EVSE wallplug.

There have been a few posts which condemn unplugging the EVSE.

I'm confused: :? does the OBC signaling back to the EVSE either do a zero-crossing timing for power cutoff or some other scheme which gradually removes power and avoids the inevitable inductive voltage spike on the input circuitry?

 

[coulomb]

There are four ways that the car turns off charging -

1. When it is fully charged, the OBC sends a shutoff signal to the EVSE which stops the power flow

By this time, the charge current has reduced to a small fraction of rated current. But the principle is still the same (see below).

2. When timed using the Remote, presumably the Remote receiver sends a signal to the OBC which then sends the shutoff signal to the EVSE

My understanding is that the EVCU (or whatever controls charging) disconnects a relay (K1 below) that was putting a 1.3 kΩ (R2 below) load on the pilot signal, so the pilot signal jumps from a peak of +6 V (charging) to a peak of +9 V (not charging, vehicle ready). (The diode ensures that the pilot signal always troughs at -12 V).

From Wikipedia's SAE J1772 page.

This presumably triggers the EVSE to open its contactor (not shown above), which is the equivalent of pulling the plug. But I also assume that before the EVCU drops the relay, it will ramp the charging power down, say over about 10 mains cycles (~170 ms, doesn't have to be a whole number of mains cycles). This allows the energy in the PFC inductor to fully flow to the battery, and the mains current is nearly zero by the end. So now after the relay is dropped and the EVSE drops the mains contactor, nothing bad happens.

3. Depressing the pushbutton on the J1772 connector sends the shutoff signal to the EVSE

The proximity signal (from the pushbutton) usually doesn't connect to the EVSE (it seems that some do, most don't). The pushing of the button changes some voltages at the vehicle, which is detected by the EVCU, which again ramps down the charge current and opens the relay.

4. Input power is removed from the EVSE, either by the use of a timer or simply pulling out the EVSE wallplug.

In this case, the EVCU gets no warning, and can't ramp the charging power down. The PFC inductor stores considerable energy, and now one end of it is effectively open circuit. The inductor tries to keep the current flowing, because that's what inductors do, and it will attempt to do this by creating a high (kilovolt) potential at the now open circuited end. Usually, there are parts that are designed to absorb this pulse.

There have been a few posts which condemn unplugging the EVSE.

I'm probably the major proponent of this condemnation, and my concern is based on experience with a completely different EV charger, so this should be taken with appropriate scepticism.

I'm confused: :? does the OBC signaling back to the EVSE either do a zero-crossing timing for power cutoff or some other scheme which gradually removes power and avoids the inevitable inductive voltage spike on the input circuitry?

As per the above, my understanding is that it gradually reduces charging power, though over a time scale that is fairly short by human perception.

As I wrote the above, I note:

• As I wrote above, there are usually parts that are designed to absorb the "pull the plug" transient. In the case of the Elcon/TC chargers I'm familiar with, these components appear to be rather under-sized, so they tend to fail in such a way that they are ineffective, but still allow the charger to operate. One would hope that the iMiEV's OBC would have more robust parts.
• Assuming these parts designed to absorb the transient have failed, then in order for the small blue capacitors to fail, the transient has to be transferred from the input stage through the IGBTs, transformer, and rectifier to the blue capacitors. I'm unsure how believable this transfer is.

[ Edit: "voltage" -> "pilot signal" ]
[ Edit: "±6 V" -> "peak of +6 V"; added sentence about the troughs always being -12 V. ]

 

[JoeS]

Coulomb, thank you very much for your detailed explanation with the schematic - I should have reviewed the J1772 interface myself before posting. The concept of the J1772 switch triggering the vehicle controller to first ramp down the charging power had escaped me.

Presumably, using the Remote does the same thing.

As a result, we're now going to become far more proactive in using the Remote for the timing of our charging and I'll eliminate use of the mechanical timers completely.

Off-topic - Coulomb, as a personal aside, my wife and I have very fond memories of Brisbane as we stayed on the boat in Manly cumulatively for about a year and our weekly treat was a daylong visit into the city by train, culminating in dinner and a movie and especially the special-occasion fireworks at South Bank. Australia Day happens to be our Anniversary.

 

[Phximiev]

Coulomb, thank you very much for your detailed explanation with the schematic - I should have reviewed the J1772 interface myself before posting. The concept of the J1772 switch triggering the vehicle controller to first ramp down the charging power had escaped me.
. . . As a result, we're now going to become far more proactive in using the Remote for the timing of our charging and I'll eliminate use of the mechanical timers completely. . .

"Hear, Hear" on the kudos and explanation. In fact, I went out and checked the portable chargers that we have and use (Mitsu, Volt, and Turbocord) and "tested" the Volt (wife ran off with the Mitsu!). Pressing the button alone without withdrawing the J1772 Turbocord caused the Volt to shut off charging and noticeably turned the charging light on the dash off. I would imagine that that such voltage drop is measurable on Canion or other programs?

So on the other side of the argument, is it fair to say that if one simply plugs in the J1772 in without engaging the button, that the charging process will not start?

Does this perhaps explain our experience why on occasion when out at a public charger where the button doesn't engage, that charging fails to start? (see Plugshare activity feed for a number of other, similar, examples)

 

[PV1]

So on the other side of the argument, is it fair to say that if one simply plugs in the J1772 in without engaging the button, that the charging process will not start?

Does this perhaps explain our experience why on occasion when out at a public charger where the button doesn't engage, that charging fails to start? (see Plugshare activity feed for a number of other, similar, examples)

Yes, if that button doesn't release (connector latch clicking into place), the EVSE will read a car connected but not ready to charge, and therefore won't start charging. Make sure that J1772 connector is all the way in and the lock fully engaged. On some charging stations, pushing the connector straight in until it stops and then lifting it slightly is required to latch the connector.

Given that the latch button sends the shutdown signal, does anyone have the capacity to measure charge current reduction during the disconnect process? What I mean is, can we measure if the charger ramps down when the latch button is pressed or does it simply hard stop charging? With my Bolt, pressing the button causes the EVSE contactor to drop out almost immediately, seemingly just under half a second. I'm curious if that is enough time for a graceful shutdown of the OBC or if using the remote to stop charging would be kinder to the hardware as that is totally controlled by the car.

 

[coulomb]

Pressing the button alone without withdrawing the J1772 Turbocord caused the Volt to shut off charging and noticeably turned the charging light on the dash off.

Did you hear the contactor in the EVSE drop out as well? Was there a short but discernible (less than a second) delay between pressing the button and the contactor dropping out?

 

[JoeS]

Depressing the button on the J1772 connector while its plugged into the car results in a seemingly-instantaneous relay click in the EVSE. The resolution of my power monitor isn't good enough - need a scope.

Edit: Update.

This morning I paid attention and I could swear that the sequence was (the car still had not run out of its Remote-set time):

Buttonpush --> click in car --> click in EVSE (Mitsu EVSE 120vac)

Just went out to confirm this, but this time I have the car plugged into my eMotorWerks EVSE (240vac) as I'm about to go on a longer trip and am topping up -

Buttonpush --> click in EVSE --> multiple clicks in car

Anyone else?

In any case, the time delays are noticeable. Will pay attention each time I unplug and will update if I notice anything different.

 

[Phximiev]

Pressing the button alone without withdrawing the J1772 Turbocord caused the Volt to shut off charging and noticeably turned the charging light on the dash off.

Did you hear the contactor in the EVSE drop out as well? Was there a short but discernible (less than a second) delay between pressing the button and the contactor dropping out?

I don’t recall hearing anything but I will check again later as well as the time lapse.

 

[Phximiev]

Pressing the button alone without withdrawing the J1772 Turbocord caused the Volt to shut off charging and noticeably turned the charging light on the dash off.

Did you hear the contactor in the EVSE drop out as well? Was there a short but discernible (less than a second) delay between pressing the button and the contactor dropping out?

I don’t recall hearing anything but I will check again later as well as the time lapse.

The contractor clicked and there was a delay, probably no more than a second.

Also with the Volt 120v charger, I toggled the switch several times with the same result: (1) button off, contactor clicked, Volt lights off, (2) button on, contactor clicked, Volt lights on.

 

[Aerowhatt]

Mitsu tested things for the way that they suggest they be used. So as long as one is starting and stopping charging in the ways supplied and suggested in the manual, then things should be alright. There are externals like line surges and spikes, power outages, etc., that the system must absorb rarely as well. It's very good idea to have a surge suppressor tied into your main breaker panel. If one charges with 120V only, then a good quality plug in line surge suppressor would be advisable. The kind one would run power through to protect a desktop computer for instance. Most of these will still provide power after the surge suppression section has sacrificed itself clamping a big surge, or spike. I prefer one that will no longer provide power after the surge suppression has been defeated. That way it's obvious that in needs replaced.

The one thing Mitsu and the charger manufacturer did not coordinate well on (IMO) is the mounting location of the OBC and how well (or poorly) it could convection cool the case while functioning. It does heat up less running on 120V, but it also runs a lot longer too. So, I think there, we may be trading heat stress for MTBF (Mean Time Between Failure) component issues. I'm making sure that the chargers on our cars get plenty of active ventilation while charging at 240V. Charging is (has been) always stopped automatically by the cars systems, or by using the push button on the EVSE nozzle as recommended in the manual. I was toying with the idea of offering up kits for actively cooling the OBC but there wasn't sufficient interest to make it worthwhile last summer when the failures started showing up.

For the 12 volt battery. I have found that regular testing is essential for all uses. Not just cranking amps but the longer, more difficult to do capacity testing. IMO a 12 volt car battery that has less than 75% of its new reserve capacity should be replaced especially on an IMiev. Shorted cells, etc., almost always occur on batteries that would test below this threshold. My 2014 has had a new battery in it for about a year now since the original failed its biannual test. At the time, that battery was only a bit over 3 years old . Another chemistry is tempting and there are some affordable options out there. I'm not sure though, that an alternate chemistry ameliorates the problem. Any battery chemistry can have a sudden weakness, or cell failure. At least we know lead acid and it's proclivities well.

Aerowhatt

 

[kiev]

Recently i made an effort to notice the J 1772 push-button sequence on my Bllink Level 2 (240 vac) EV SE.

It occurs almost instantly, but after a very brief delay the EV SE relay opens first, then about 1 second or less later, the contactors in the car can be heard to open.

[edit]
i repeated the observation using the OEM level 1 (120 vac) EV SE, and this time the delays were in the same sequence but were noticeably longer.

 

[bradleydavidgood777]

I've been trying to follow this thread but haven't kept up to it all, and much of it is over my head so I get tired of reading.

I am mostly trying to understand the current theories for failures and if what can be done to hopefully avoid this issue.

This is what I understand...and please, correct me or add to this.

1. Do not let your 12V battery get low or old.
2. Do not cut power to your EVSE while charging. Wait until the charge is finished. Then remove the plug from the car first. If you can't wait for the charge to finish, remove the plug from the car while charging first, instead of cutting power from the EVSE or utility power side.

Let me know if I missed anything.

Questions:

Should I make it routine maintenance to have my 12V battery checked at inspection time? And if so, what kind of reading should I be looking for to justify replacing it?

Should I consider adding some kind of protection before this failure point? If someone here knew what to do and how to do it, I would gladly get in line for this upgrade. I'm pretty handy so I think I could probably remove parts if it is not that difficult and does not require special equipment. Then could ship parts to someone and they could do the upgrade and send them back. Has anyone considered this? Could be a good way to make some extra cash and would also be a great contribution to this group. Think about how great it would be to tell a new owner that they could avoid this issue by getting the upgrade.

Thanks

 

[Don]

I believe it's been determined that the primary cause of the On Board Charger failures (the $4K unit that charges the high voltage traction battery) is that it overheats. It's water cooled, but the cooling system doesn't remove much heat and on a warm day, it gets really hot inside. Kiev took many temperature readings which pretty much proved Mitsu's cooling system has major flaws

Several of us tried using fans on hot days to carry away some of the excess heat. Since the way we use our two cars never requires fast charging, I quit using 240 Volt L2 charging completely. In fact, 90% of the time for the past 8 months or so I've been using the stock 8 amp, 120 volt Panasonic EVSE for nearly all my charging. The other times I've been using my modified 12 amp 120 volt EVSE. The OBC doesn't get nearly as hot charging on L1 as it does using L2

Don